Method for coding integer supporting diverse frame sizes and codec implementing the method
AssigneeSamsung Electronics Co., Ltd.
InventorKi-Seok Chang
Life of this patent
Estimating the term…
Abstract
An integer coding method supporting diverse frame sizes and a coder-decoder (CODEC) implementing the method are provided. Maximum sample values are detected from a plurality of frames having different numbers of samples. The detected maximum sample values are normalized on the basis of a maximum sample value of the largest frame. A maximum sample value is detected from the normalized sample values. Multiples are determined according to the size of each frame on the basis of the detected maximum sample value and a sample value of each frame is multiplied by the multiples to generate an integer code.
Claims7
An integer coding method of a digital signal processing (DSP) coder-decoder (CODEC) for coding or decoding an audio or video signal, the method comprising:
- detecting maximum sample values from a plurality of frames having different numbers of samples;
- normalizing the detected maximum sample values on the basis of a maximum sample value of a largest frame having a greatest number of samples;
- detecting a maximum normalized sample value from the normalized sample values;
- and
- determining multiples according to a size of each frame on the basis of the detected maximum normalized sample value and multiplying a sample value of each frame by the multiples to generate an integer code.
A coder-decoder (CODEC) for coding or decoding audio or video data in units of frames, the CODEC comprising:
- an inverse quantization unit configured to inversely quantize samples in the form of a variable length code (VLC) and generate samples having real number values;
- an integer coding unit configured to determine multiples according to a size of a frame on the basis of a maximum sample value and an effective bit for samples in units of frames inversely quantized by the inverse quantization unit, and multiply the inversely quantized samples in units of frames by the multiples to perform an integer operation of the samples;
- and
- an integer inverse transform unit configured to perform an inverse Fourier transform operation on the samples multiplied by the integer coding unit to transform the samples into integers.
Dependencies are inferred from the claim text.